[Merged by Bors] - Fix all_tuples macro for non-0/1 starts

[james7132]

Objective

all_tuples panics when the start count is set to anything other than 0 or 1. Fix this bug.

Solution

Originally part of #2381, this PR fixes the slice indexing used by the proc macro.

[DJMcNab]

The change in itself is fine, although the motivation is unclear. I don't think we want all_tuples to be part of our public API, and obviously no internal uses run into this case.

[james7132]

Ran into it as a part of #2381, where it was explicitly used in a case where the start for the macro was set at 2 to avoid confusing duplicate cases in the API.

This has since been made into an ecosystem crate (bevy_system_graph) instead of being merged into bevy_ecs, which is now as an external consumer of the macro.

[DJMcNab]

I just saw that all_tuples is re-exported at the root of bevy_ecs. I agree then that using it publicly is 'fair'. However, I don't think we should be exporting it; evidence in favour of that is that it has no docs.

Since it is useful outside of our internal use, all_tuples should probably be spun out into its own crate.

[mockersf]

I just saw that all_tuples is re-exported at the root of bevy_ecs. I agree then that using it publicly is 'fair'. However, I don't think we should be exporting it; evidence in favour of that is that it has no docs.

The macro is used in bevy_render so it has to be public...

[cart]

bors r+

[cart]

My preference is that we leave this public and move it to bevy_utils. I don't think it deserves its own crate.

[bors]

Pull request successfully merged into main.

Build succeeded:

• build-android
• build (nightly, ubuntu-latest)
• build (stable, macos-latest)
• build (stable, ubuntu-latest)
• build (stable, windows-latest)
• build-wasm (nightly, ubuntu-latest)
• build-wasm (stable, ubuntu-latest)
• check-benches
• check-doc
• check-markdown-links
• check-missing-examples-in-docs
• check-unused-dependencies
• ci
• markdownlint
• run-examples